Effectiveness and safety of bedaquiline-containing regimens in the treatment ofMDR- and XDR-TB: a multicentre study

Sergey E. Borisov1,50, Keertan Dheda2,50, Martin Enwerem3,50,Rodolfo Romero Leyet4,50, Lia D’Ambrosio5,6,50, Rosella Centis5,50,Giovanni Sotgiu 7,50, Simon Tiberi8,9,50, Jan-Willem Alffenaar10,50,Andrey Maryandyshev11,50, Evgeny Belilovski1,50, Shashank Ganatra12,50,Alena Skrahina13,50, Onno Akkerman14,15, Alena Aleksa16, Rohit Amale12,Janina Artsukevich16, Judith Bruchfeld17, Jose A. Caminero18,19,Isabel Carpena Martinez20, Luigi Codecasa21, Margareth Dalcolmo22,Justin Denholm23, Paul Douglas24, Raquel Duarte25, Aliasgar Esmail26,Mohammed Fadul26, Alexey Filippov1, Lina Davies Forsman17, Mina Gaga27,Julia-Amaranta Garcia-Fuertes28, José-María García-García 29,Gina Gualano30, Jerker Jonsson31, Heinke Kunst9, Jillian S. Lau32,Barbara Lazaro Mastrapa33, Jorge Lazaro Teran Troya33, Selene Manga34,Katerina Manika35, Pablo González Montaner36, Jai Mullerpattan12,Suzette Oelofse26, Martina Ortelli37, Domingo Juan Palmero36,Fabrizio Palmieri30, Antonella Papalia38, Apostolos Papavasileiou39,Marie-Christine Payen40, Emanuele Pontali41, Carlos Robalo Cordeiro42,Laura Saderi7, Tsetan Dorji Sadutshang43, Tatsiana Sanukevich16,Varvara Solodovnikova13, Antonio Spanevello44,45, Sonam Topgyal43,Federica Toscanini46, Adrian R. Tramontana47, Zarir Farokh Udwadia12,Pietro Viggiani38, Veronica White48, Alimuddin Zumla49 and GiovanniBattista Migliori 5,50

@ERSpublicationsBedaquiline is safe and effective in treating MDR- and XDR-TB patients http://ow.ly/6MWK30adHkw

Cite this article as: Borisov SE, Dheda K, Enwerem M, et al. Effectiveness and safety of bedaquiline-containing regimens in the treatment of MDR- and XDR-TB: a multicentre study. Eur Respir J 2017; 49:1700387 [https://doi.org/10.1183/13993003.00387-2017].

This article has supplementary material available from erj.ersjournals.com

Earn CME accreditation by answering questions about this article. You will find these at erj.ersjournals.com/journal/cme

Received: Feb 23 2017 | Accepted after revision: March 16 2017

Conflict of interest: Disclosures can be found alongside this article at erj.ersjournals.com

The authors alone are responsible for the views expressed in this publication and they do not necessarily represent thedecisions and policies of their institutions.

Copyright ©ERS 2017

https://doi.org/10.1183/13993003.00387-2017 Eur Respir J 2017; 49: 1700387

ORIGINAL ARTICLETUBERCULOSIS

ABSTRACT Large studies on bedaquiline used to treat multidrug-resistant (MDR-) and extensivelydrug-resistant tuberculosis (XDR-TB) are lacking. This study aimed to evaluate the safety and effectivenessof bedaquiline-containing regimens in a large, retrospective, observational study conducted in 25 centresand 15 countries in five continents.

428 culture-confirmed MDR-TB cases were analysed (61.5% male; 22.1% HIV-positive, 45.6% XDR-TB). MDR-TB cases were admitted to hospital for a median (interquartile range (IQR)) 179 (92–280) daysand exposed to bedaquiline for 168 (86–180) days. Treatment regimens included, among others, linezolid,moxifloxacin, clofazimine and carbapenems (82.0%, 58.4%, 52.6% and 15.3% of cases, respectively).

Sputum smear and culture conversion rates in MDR-TB cases were 63.6% and 30.1%, respectively at30 days, 81.1% and 56.7%, respectively at 60 days; 85.5% and 80.5%, respectively at 90 days and 88.7% and91.2%, respectively at the end of treatment. The median (IQR) time to smear and culture conversion was34 (30–60) days and 60 (33–90) days. Out of 247 culture-confirmed MDR-TB cases completing treatment,71.3% achieved success (62.4% cured; 8.9% completed treatment), 13.4% died, 7.3% defaulted and 7.7%failed. Bedaquiline was interrupted due to adverse events in 5.8% of cases. A single case died, havingelectrocardiographic abnormalities that were probably non-bedaquiline related.

Bedaquiline-containing regimens achieved high conversion and success rates under differentnonexperimental conditions.

Affiliations: 1Moscow Research and Clinical Center for TB Control, Moscow Government’s Health Department,Moscow, Russian Federation. 2UCT Lung Institute, Division of Pulmonology, University of Cape Town, CapeTown, South Africa. 3Amity Health Consortium, Country Club Estate, Johannesburg, South Africa. 4ClinicalUnit, District Clinical Specialist Team, Springbok, South Africa. 5World Health Organization CollaboratingCentre for Tuberculosis and Lung Diseases, Maugeri Care and Research Institute, Tradate, Italy. 6PublicHealth Consulting Group, Lugano, Switzerland. 7Clinical Epidemiology and Medical Statistics Unit, Dept ofBiomedical Sciences, University of Sassari, Sassari, Italy. 8Division of Infection, Royal London Hospital, BartsHealth NHS Trust, London, UK. 9Blizard Institute, Barts and The London School of Medicine and Dentistry,Queen Mary University of London, London, UK. 10University of Groningen, University Medical CenterGroningen, Dept of Clinical Pharmacy and Pharmacology, Groningen, The Netherlands. 11Northern StateMedical University, Arkhangelsk, Russian Federation. 12Dept of Respiratory Medicine, P.D. Hinduja NationalHospital and MRC, Mumbai, India. 13Republican Research and Practical Centre for Pulmonology andTuberculosis, Minsk, Belarus. 14University of Groningen, University Medical Center Groningen, TuberculosisCenter Beatrixoord, Haren, The Netherlands. 15University of Groningen, University Medical Center Groningen,Dept of Pulmonary Diseases and Tuberculosis, Groningen, The Netherlands. 16Dept of Phthisiology, GrodnoState Medical University, GRCC “Phthisiology”, Grodno, Belarus. 17Unit of Infectious Diseases, Dept ofMedicine, Solna, Karolinska Institute, Dept of Infectious Diseases, Karolinska University Hospital, Stockholm,Sweden. 18Pneumology Dept, Hospital General de Gran Canaria “Dr Negrin”, Las Palmas de Gran Canaria,Spain. 19MDR-TB Unit, Tuberculosis Division, International Union against Tuberculosis and Lung Disease (TheUnion), Paris, France. 20General University Hospital Morales Meseguer, Murcia, Spain. 21TB Reference Centre,Villa Marelli Institute/Niguarda Hospital, Milan, Italy. 22Hélio Fraga Reference Center, Fiocruz/MoH, Rio deJaneiro, Brazil. 23Victorian Tuberculosis Program, Melbourne Health, Dept of Microbiology and Immunology,University of Melbourne, Peter Doherty Institute for Infection and Immunity, Melbourne, Australia. 24HealthPolicy and Performance Branch, Health Services and Policy Division, Dept of Immigration and BorderProtection, Sydney, Australia. 25National Reference Centre for MDR-TB, Hospital Centre Vila Nova de Gaia,Dept of Pneumology, Public Health Science and Medical Education Department, Faculty of Medicine,University of Porto, Porto, Portugal. 26UCT Lung Institute, Lung Infection and Immunity Unit, Division ofPulmonology, Dept of Medicine, University of Cape Town, Groote Schuur Hospital, Cape Town, South Africa.277th Respiratory Medicine Dept, Athens Chest Hospital, Athens, Greece. 28Bronchiectasis Unit, RespiratoryDept, Hospital Universitario Araba, Vitoria-Gasteiz, Spain. 29Tuberculosis Research Programme, SEPAR,Barcelona, Spain. 30Respiratory Infectious Diseases Unit, National Institute for Infectious Diseases“L. Spallanzani”, IRCCS, Rome, Italy. 31National TB Surveillance Unit, Public Health Agency, Stockholm,Sweden. 32Dept of Infectious Diseases, Box Hill Hospital, Victoria, Australia. 33Harry Surtie Hospital, Upington,South Africa. 34Dept of Infectious Diseases, University National San Antonio Abad Cusco, Cusco, Perù.35Pulmonary Dept, ‘G. Papanikolaou’ Hospital, Aristotle University, Thessaloniki, Greece. 36PulmonologyDivision, Municipal Hospital F. J. Munĩz, Buenos Aires, Argentina. 37Pneumology Dept, University of Insubria,Varese, Italy. 38AOVV Eugenio Morelli Hospital, Reference Hospital for MDR and HIV-TB, Sondalo, Italy.39MDR-TB Unit, Athens Chest Hospital, Ministry of Health, Athens, Greece. 40Division of Infectious Diseases,CHU Saint-Pierre, Université Libre de Bruxelles (ULB), Brussels, Belgium. 41Dept of Infectious Diseases,Galliera Hospital, Genoa, Italy. 42Coimbra Medical School, Pneumology Dept, Coimbra University Hospital,Coimbra, Portugal. 43Delek Hospital, Dharamshala, India. 44Pneumology Dept, Maugeri Care and ResearchInstitute, Tradate, Italy. 45Dept of Clinical and Experimental Medicine, University of Insubria, Varese, Italy.46University Hospital San Martino, Care and Research Institute, National Institute for Cancer Research, Genoa,Italy. 47Dept of Infectious Diseases, Western Hospital, Victoria, Australia. 48Dept of Respiratory Medicine,Barts Healthcare NHS Trust, London, UK. 49Division of Infection and Immunity, University College London andNIHR Biomedical Research Centre, UCL Hospitals NHS Foundation Trust, London, UK. 50These authorscontributed equally.

Correspondence: G.B. Migliori, World Health Organization Collaborating Centre for Tuberculosis and LungDiseases, Maugeri, Care and Research Institute, Via Roncaccio 16, 21049, Tradate, Italy.E-mail: giovannibattista.migliori@icsmaugeri.it

https://doi.org/10.1183/13993003.00387-2017 2

TUBERCULOSIS | S.E. BORISOV ET AL.

Introduction480000 cases of multidrug-resistant tuberculosis (MDR-TB) and 100000 cases of rifampicin-resistant(RR)-TB eligible for MDR-TB treatment were estimated by the World Health Organization (WHO) tohave occurred in 2015, with 250000 deaths [1]. Over half of the estimated MDR-TB cases occurred inIndia, China, the Russian Federation and the other former Soviet Union countries, as well as in SouthAfrica [1]. Globally, ∼10% of the MDR-TB strains meet the criteria defining extensively drug-resistant(XDR)-TB (resistance to any fluoroquinolone and at least one second-line injectable drug) [1, 2].

Treatment for MDR- and XDR-TB is long, expensive and characterised by a high rate of adverse events[3–15]. The main difficulty is the identification of at least four active drugs in order to design an effectiveregimen [3, 5, 7, 10, 11].

The prior stepwise approach based on the hierarchical use of first- and second-line anti-TB drugsclassified into five groups has been recently modified by the WHO. The new classification includes fourgroups of drugs (A: fluoroquinolones; B: second-line injectable agents; C: other core second-line agents;and D: add-on agents, subdivided into the subgroups D1, D2 and D3) [5].

Two newly available drugs, delamanid [12–14] and bedaquiline [15–22], together with some repurposeddrugs (linezolid [7, 23–26], carbapenems [27–31] and clofazimine [32–34], among others [35, 36]) arepresently pivotal in ongoing scientific discussions.

The information available today on bedaquiline is still limited to phase 2 studies and small individualcohorts treated under clinical trial conditions, the largest study not exceeding 233 patients [15–21, 37]. Inparticular, at this present time, no study of size informs us on the effectiveness, safety and tolerability ofbedaquiline, in different continents and programmatic conditions.

Given the concerns around the adverse events of bedaquiline (particularly QT prolongation, potentially athighest risk when added to the fluoroquinolones moxifloxacin, levofloxacin and oflofloxacin, clofazimine,delamanid and methadone, among others), additional evidence on its safety, apart from that provided bythe registration trials, is urgently needed [38–40].

Recently, TB reference centres belonging to the International Bedaquiline Study Group (a networkmerging the centres belonging to the International Carbapenems Study Group (ICSG) [29–31] and theEuropean Respiratory Society (ERS)/Asociación Latinoamericana de Tórax (ALAT) and the BrazilianSociety of Respiratory Medicine collaborative projects [41, 42], coordinated by the ERS TB collaboratingcentre) conducted an observational study on the therapeutic contribution of bedaquiline within abackground regimen (as per WHO guidelines) when treating MDR- and XDR-TB cases.

The aim of the present study is to evaluate the safety, tolerability and effectiveness of bedaquiline withinbackground regimens in a large multicentre cohort of MDR- and XDR-TB patients treated under differentconditions (programmatic, expanded access, but not recruited for experimental studies).

Material and methodsThe methodological approach adopted for this study is similar to that described in previous ICSG studies[29–31]. 25 MDR-TB reference centres located in 15 countries in Africa, Asia, Western and EasternEurope, Oceania and Southern America (Argentina, Australia (Victoria State), Belarus, Belgium, Greece,India, Italy, the Netherlands, Peru, Portugal, Russian Federation (Arkhangelsk and Moscow city), SouthAfrica, Spain, Sweden and United Kingdom) retrospectively collected data on RR-TB andculture-confirmed MDR-TB patients aged ⩾15 years (figure 1 and online supplementary table S1).

An MDR-TB case was defined as an individual with TB disease caused by Mycobacterium tuberculosisstrains phenotypically resistant to at least isoniazid and rifampicin. RR-TB cases were those diagnosedusing Xpert MTB/RIF (Cepheid, Sunnyvale, CA, USA) complemented by line probe assays. XDR-TB caseswere those whose disease was due to MDR M. tuberculosis strains with additional resistance to anyfluoroquinolone and one of the second-line injectable drugs (i.e. amikacin, capreomycin and kanamycin)[1, 5].

Patients starting their treatment between January 1, 2008 and August 30, 2016 were consecutively enrolledbased on their exposure to bedaquiline during the intensive and continuation phases. We included in thestudy only patients exposed to bedaquiline-containing regimens, without considering the date of exposure.The condition was that bedaquiline had to be added within the design of the last treatment course.

An individualised TB regimen was administered following the results of drug susceptibility testing (DST)performed by externally quality-assured laboratories [29–31]. RR-TB cases were managed according to thenational guidelines in force in their respective countries (e.g. South Africa) [43].

https://doi.org/10.1183/13993003.00387-2017 3

TUBERCULOSIS | S.E. BORISOV ET AL.

Physicians were free to prescribe the accompanying anti-TB treatment to obtain the best possible regimenin their setting and, consequently, no specific protocol or method beyond local guidelines was followed.

Bedaquiline was administered at the recommended dose of 400 mg once a day for 14 days, then 200 mgthree times a week for 22 weeks. It was prescribed under programmatic conditions, expanded access andcompassionate use, but not under the experimental protocols of clinical trials. Adverse events wereattributed to bedaquiline or to another particular drug by local attending physicians and local teams (withsupport of a local TB consilium, where available), without any standardised protocols or documented levelof ascertainment of the causal association. A standardised classification for the severity of the adverseevents was not used.

A standardised ad hoc e-form was used to collect epidemiological (i.e. age; place of birth and residence;sex; and migrant status from a TB high-incidence country), clinical (i.e. cardiac and thyroid disorders;HIV testing; HIV infection status; administration of HIV drugs; previous TB diagnosis and treatment;previous treatment outcomes; radiological findings; TB therapy and related adverse events; duration ofexposure to bedaquiline, delamanid, linezolid or carbapenems; adjuvant surgery; sputum smear andculture positivity at baseline; and during treatment at 30, 60 and 90 days, time to sputum smear andculture conversion; WHO treatment outcomes; and duration of hospital stay) and microbiological (i.e.DST results) data from medical records. Culture conversion rates were calculated, including in thenumerator the patients with a negative culture, and in the denominator the patients completing treatmentincluding failures, but excluding death and default outcomes.

As the latest WHO outcome definitions were introduced in 2013 and our first cohort was treated in 2008,we kept the outcome “default” instead of “lost to follow-up” [5].

With regard to the analysis of treatment outcomes, a subanalysis of MDR-TB patients who started theirtreatment before December 31, 2014 was performed in order to include those who could fit the definitionof treatment completion and cure according to the WHO guidelines on MDR-TB treatment [5].

A homogeneous and standardised approach to DST was not possible in all the clinical centres. Thepercentages reported for resistance have been calculated according to the number of cases tested for(denominator) and the number of cases resistant to (numerator) a specific drug.

Ethical approval for the retrospective collection of clinical data was obtained by the coordinating centre. Thecentres involved in the study requested the authorisation for the treatment of their patients according tonational legislation, and, accordingly, a consent form was signed by the patient and the attending physician.

The map in figure 1 was created using the rworldmap package (www.r-project.org) [44].

Qualitative and quantitative variables were summarised using percentages and medians (interquartile range(IQR)). Chi-squared or Fisher exact tests were used to compare qualitative variables, and the Mann–Whitney test was used to statistically compare quantitative variables.

FIGURE 1 Global distribution of multidrug-resistant and extensively drug-resistant tuberculosis cases treatedwith regimens containing bedaquiline, 2009–2016. n=428 cases in 25 study sites (online supplementarytable S1).

https://doi.org/10.1183/13993003.00387-2017 4

TUBERCULOSIS | S.E. BORISOV ET AL.

A p-value of <0.05 was considered statistically significant. Statistical computations were performed usingStata 13.0 (StataCorp, College Station, TX, USA).

ResultsPatient characteristicsDemographic, epidemiological and clinical characteristics of the patients are summarised in table 1 andfigure 1; safety and tolerability information on bedaquiline-containing regimens in table 2; and treatmentoutcome results in tables 3 and 4. 428 culture-confirmed MDR-TB patients were recruited (figure 1 andonline supplementary table S1). The most prevalent sex in the cohort was male (n=263, 61.5%) and themedian (IQR) age was 35 (27–44) years. The characteristics of the 428 culture-confirmed MDR-TBpatients are summarised in table 1. There were 45 (10.5%) migrants from countries with high TBincidence to those with low TB incidence.

The proportion of HIV co-infected patients was 22.1%; their median (IQR) CD4 cell count was 269 (168–470) cells·mm−3 and the majority (n=92, 97.9%) received antiretroviral therapy. HIV co-infectionprevalence in Africa, Eastern Europe and remaining settings was 88 (46.3%) out of 190, 0 (0%) out of 150and 6 (7.1%) out of 85, respectively.

Pulmonary TB was diagnosed in 426 (99.5%) out of 428 cases; the extrapulmonary locations beingabdominal and the nervous system (n=2). The percentages of sputum smear- and culture-positive caseswere 72.1% and 98.4%, respectively.

Fewer than half the patients were affected by XDR-TB (195/428, 45.6%), with a median (IQR) number ofdrug resistances of 3 (1–5). Overall, 334 (78.0%) out of 428 cases had been treated previously for TB.

The prevalence of drug resistance was as follows: streptomycin 185 (94.4%), pyrazinamide 145 (70.4%),fluoroquinolones 267 (64.5%), amikacin 131 (44.4%), capreomycin 127 (41.6%), kanamycin 179 (59.3%),ethionamide 135 (59.7%), para-aminosalicylic acid 70 (35.7%), linezolid 4 (10.5%), ethambutol 186(77.5%) and cycloserine 20 (12.3%).

Treatment regimens included linezolid (82.0%), clofazimine (52.6%), moxifloxacin (58.4%), second-lineinjectables (45.8%), carbapenems (15.3%) and ofloxacin (1.6%).

The median (IQR) range of the administrative delay in procuring bedaquiline was 8 (0–60) days.

Patients were exposed to bedaquiline for a median (IQR) of 168 (86–180) days (table 2). Five (1.2%)patients underwent treatment with both delamanid and bedaquiline. Adjuvant surgical therapy wasperformed in 55 (13%) cases.

The median (IQR range) treatment duration in the cohort was 18 (10–22) months.

Adverse eventsThe majority of the adverse events reported were nausea, peripheral neuropathy and otovestibular toxicity.Adverse events potentially attributed to bedaquiline were reported in 80 (19.4%) out of 413 cases wherethis information was provided (table 2).

In particular, 51 (11.9%) out of 428 patients discontinued bedaquiline (25 (5.8%) of whom reportedadverse events); of these 26 (51%) did so permanently.

Although we do not have the exact information on how many patients interrupted bedaquiline due toFridericia-corrected QT interval (QTcF) increase, 24 (9.7%) out of 247 patients experienced QTcFprolongation >500 ms.

One patient was started on bedaquiline with a baseline QTcF of 553 ms, which then decreased to 536 msat week 4 and 554 ms at week 8. A second patient, with a baseline QTcF of 352 ms, had a transientincrease (510 ms at week 3) followed by a decline (358 ms at week 4); she was palliated at home as per herrequest. She had had no arrhythmias at any point and no evidence to suggest that her death was related toa cardiovascular event. She achieved sputum conversion (smear and culture) prior to her death.

Figure 2 summarises the median values of the QTcF interval and its temporal trends in the cohort overthe first 12 weeks of treatment.

The median (IQR) exposure to bedaquiline among the 26 patients who permanently interrupted treatmentwas 69 (27.5–135) days, and 85.5 (44.3–160) days in the 33 patients who died.

Out of 33 patients who died, we have QT information on 19 (57.6%), and only one patient had a baselineQT >500 ms. During follow-up none of those with a normal QT at baseline showed an increase >500 ms.

https://doi.org/10.1183/13993003.00387-2017 5

TUBERCULOSIS | S.E. BORISOV ET AL.

TABLE 1 Demographic, epidemiological and clinical characteristics of 428 culture-confirmedmultidrug-resistant tuberculosis (MDR-TB) patients exposed to bedaquiline-containingregimens

Age years 35 (27–44)Male 263/428 (61.5)Body weight at admission kg 56 (47–65)Height cm 169 (160–176)Migrant 45/428 (10.5)Employed 110/413 (26.6)Prisoner 23/413 (5.6)Pregnant 1/133 (0.8)Thyroid disease 12/413 (2.9)Heart disease 40/318 (12.6)Pre-existing ECG abnormality 13/318 (4.1)Alcoholism 132/363 (36.4)Drug abuse 75/413 (18.2)Methadone use 0/318 (0.0)Diabetes 26/413 (6.3)HIV infection testing 425/428 (99.3)HIV-positive 94/425 (22.1)CD4 count at baseline cells·mm-3 269 (168–470)Nadir CD4+ cell count cells·mm-3 222 (160–402)Antiretroviral therapy exposure 92/94 (97.9)Prior anti-TB therapy 334/428 (78.0)Prior courses of treatment lasting >1 month n 2 (1–2)Prior TB treatment outcomeSuccess 68/263 (25.9)Cure 41/263 (15.6)Completion 27/263 (10.3)Default 35/263 (13.3)Transfer out 20/263 (7.6)Failure 140/263 (53.2)

Pulmonary surgery 55/423 (13.0)Previous MDR-TB 169/317 (53.3)Pulmonary involvement 426/428 (99.5)Extrapulmonary involvement 11/428 (2.6)Radiology involvementCavitary lesion 81/331 (24.5)Bilateral pulmonary with cavitary lesions 147/331 (44.4)Bilateral pulmonary 56/331 (16.9)Noncavitary nonbilateral pulmonary 47/331 (14.2)

Sputum smear positive 305/423 (72.1)Sputum culture positive 421/428 (98.4)Drug resistanceStreptomycin 185/196 (94.4)Ethambutol 186/240 (77.5)Pyrazinamide 145/206 (70.4)Fluoroquinolones 267/414 (64.5)Amikacin 131/295 (44.4)Capreomycin 127/305 (41.6)Kanamycin 179/302 (59.3)Ethionamide 135/226 (59.7)Cycloserine 20/163 (12.3)Para-aminosalicylic acid 70/196 (35.7)Linezolid 4/38 (10.5)Rifabutin 32/35 (91.4)

XDR-TB 195/428 (45.6)MDR- or XDR-TB contacts 105/411 (25.6)Hospital stay days 179 (92–280)

Data are presented as median (interquartile range) or n/N (%). XDR-TB: extensively drug-resistanttuberculosis.

https://doi.org/10.1183/13993003.00387-2017 6

TUBERCULOSIS | S.E. BORISOV ET AL.

According to the information available, a single case out of 33 who died had ECG abnormalities. Duringbedaquiline treatment the patient had a single QTcF measurement >450 ms, but <500 ms. This patientreported to the primary health centre 2 days before death complaining of weakness. The coordinatingcentre suggested drawing blood for electrolyte analysis, which revealed hypokalaemia. Thus, the patientwas transferred to the district hospital where ECG showed premature ventricular complex bigeminy andthen irreversible cardiac arrest. He died after 131 days of treatment with bedaquiline. The final ECGshowed QTcF <450 ms. Thus, a relationship between bedaquiline use and fatal arrhythmia seems unlikelyfollowing revision of the clinical and ECG history of the patient.

Furthermore, 104 (29.9%) out of 348 cases treated with linezolid reported adverse events attributed to thisdrug by the attending physician; 16 (27.6%) out of 58 of them (for whom final treatment outcome wasavailable) permanently discontinued linezolid.

Sputum conversion and treatment outcomesSputum smear and culture conversion rates were 63.6% and 30.1%, respectively at 30 days; 81.1% and56.7%, respectively at 60 days; 85.5% and 80.5%, respectively at 90 days and 88.7% and 91.2%, respectivelyat the end of the MDR-TB treatment (for those completing it) respectively. The median (IQR) time tosputum smear and culture conversion was 34 (30–60) days and 60 (33–90) days, respectively (figure 3).

Out of 247 culture-confirmed MDR-TB cases that completed treatment, 71.3% achieved success (62.4%cured and 8.9% completed treatment), 13.4% died, 7.3% defaulted and 7.7% failed (table 3). The treatmentoutcomes of the 130 patients initiating their treatment before December 31, 2014 are reported in table 3.

TABLE 2 Safety and tolerability profile of bedaquiline-containing regimens in a cohort of 428culture-confirmed multidrug-resistant tuberculosis patients

Interruption of bedaquiline 51/428 (11.9)Interruption of bedaquiline due to adverse events 25/428 (5.8)Adverse events presumably due to bedaquiline 80/213 (19.4)Bedaquiline restarted if interrupted 25/69 (36.2)Total bedaquiline exposure days 168 (86–180)Creatinine >1.4×ULN 91/411 (22.1)Lipase >1.6×ULN 1/239 (0.4)ALT >3×ULN 92/413 (22.3)Bilirubin >2×ULN 47/413 (11.4)Albumin g·dL−1 36 (30–40)Potassium <3.4 or >5.6 mmol·L−1 98/412 (23.8)Magnesium <0.59 mmol·L−1 21/199 (10.6)Calcium <1.75 mmol·L−1 23/302 (7.6)Adverse effectsNausea 130/413 (31.5)Peripheral neuropathy 96/412 (23.3)Otovestibular toxicity 96/412 (23.3)Vomiting 87/411 (21.2)Anaemia 86/412 (20.9)Arthralgia 84/412 (20.4)Skin rash 63/412 (15.3)Diarrhoea 56/412 (13.6)Renal failure 47/413 (11.4)Thrombocytopenia 41/413 (9.9)Neutropenia 40/413 (9.7)Lymphocytopenia 40/413 (9.7)QT prolongation 24/248 (9.7)Hypothyroidism 38/410 (9.3)Psychiatric disorder 29/413 (7.0)Tendinopathy 18/413 (4.4)Optic neuropathy 10/413 (2.4)Deep vein thrombosis 7/412 (1.7)Pancreatitis 4/318 (1.3)Hallucinations 2/411 (0.5)Stroke 1/318 (0.3)

Data are presented as n/N (%) or median (interquartile range). ULN: upper limit of normal; ALT:alanine-transaminase.

https://doi.org/10.1183/13993003.00387-2017 7

TUBERCULOSIS | S.E. BORISOV ET AL.

TABLE 3 Treatment outcome and conversion rates of culture-confirmed multidrug-resistant tuberculosis (MDR-TB) patientsexposed to bedaquiline-containing regimens and treatment outcomes of MDR-TB patients starting their treatment withbedaquiline-containing regimens before December 31, 2014

Bedaquiline-containing regimenBedaquiline-containing regimen started

before December 31, 2014

Patients n 428 130Sputum smear conversion 30 days 206/324 (63.6)Sputum culture conversion 30 days 117/389 (30.1)Sputum smear conversion 60 days 257/317 (81.1)Sputum culture conversion 60 days 213/376 (56.7)Sputum smear conversion 90 days 265/310 (85.5)Sputum culture conversion 90 days 298/370 (80.5)Sputum smear conversion at the end of treatment 126/140 (90.0)Sputum culture conversion at the end of treatment 191/208 (91.8)Time to sputum smear conversion days 4 (30–60)Time to sputum culture conversion days 60 (33–90)Treatment outcomeSuccess 176/247 (71.3) 100/130 (76.9)Cure 154/247 (62.4) 86/130 (66.2)Completion 22/247 (8.9) 14/130 (10.8)Death 33/247 (13.4) 8/130 (6.2)Default 18/247 (7.3) 6/130 (4.6)Failure 19/247 (7.7) 16/130 (12.3)Transfer out 1/247 (0.4)

Duration of treatment after MDR-TB diagnosis months 18 (10–22)Body weight at end of treatment kg 61.9±13.9

Data are presented as n/N (%), median (interquartile range) or mean±SD, unless otherwise stated.

TABLE 4 Treatment outcomes of 428 culture-confirmed multidrug resistant (MDR-) andextensively drug-resistant tuberculosis (XDR-TB) patients exposed to bedaquiline-containingregimens in different settings and stratified by cohort

Africa Eastern Europe Other settings 2008–2009 2010–2011 2012–2014

Total cohort 113 85 49 9 121Success 73 (64.6) 65 (76.5) 38 (77.6) 5 (55.6) 95 (78.5)Cure 73 (64.6) 54 (63.5) 27 (55.1) 5 (55.6) 81 (66.9)Completion 11 (12.9) 11 (22.5) 0 (0.0) 14 (11.6)Death 27 (23.9) 3 (3.5) 3 (6.1) 2 (22.2) 6 (5.0)Default 9 (8.0) 8 (9.4) 1 (2.0) 2 (22.2) 4 (3.3)Failure 3 (2.7) 9 (10.6) 7 (14.3) 0 (0.0) 16 (13.2)Transfer out 1 (0.9)

MDR-TB 62 39 27Success 36 (58.1) 28 (71.8) 22 (81.5)Cure 36 (58.1) 20 (51.3) 11 (40.7)Completion 8 (20.5) 11 (40.7)Death 17 (27.4) 3 (7.7) 1 (3.7)Default 7 (11.3) 3 (7.7) 1 (3.7)Failure 1 (1.6) 5 (12.8) 3 (11.1)Transfer out 1 (1.6)

XDR-TB 51 46 22Success 37 (72.6) 37 (80.4) 16 (72.7)Cure 37 (72.6) 34 (73.9) 16 (72.7)Completion 3 (6.5)Death 10 (19.6) 2 (9.1)Default 2 (3.9) 5 (10.9)Failure 2 (3.9) 4 (8.7) 4 (18.2)Transfer out

Data are presented as n or n (%).

https://doi.org/10.1183/13993003.00387-2017 8

TUBERCULOSIS | S.E. BORISOV ET AL.

Sputum smear and culture conversion rates at the end of treatment were not significantly differentbetween XDR- and MDR-TB cases (p=0.73 and 0.96, respectively).

The treatment success rates for MDR- and XDR-TB patients were higher in Eastern Europe and in settingsother than in Africa; table 4 summarises the treatment outcome by cohort.

DiscussionThe aim of the present study was to retrospectively evaluate the effectiveness, safety and tolerability ofbedaquiline-containing regimens in a large observational cohort of MDR- and XDR-TB patients treatedunder different conditions (programmatic, expanded access, but not recruited for experimental studies).

To our knowledge, this is the largest study describing how safe, well tolerated and effective bedaquiline iswithin background regimens and the first in the scientific literature reporting on the nonclinical trial useof bedaquiline for the treatment of MDR- and XDR-TB patients in five continents.

In comparison with the data on individual patient data analysis (where there was a 43% success inXDR-TB) [3], in our bedaquiline-treated XDR-TB cohort the success rate was 71.3%.

In a phase 2 double blind, randomised controlled trial study by DIACON et al. [16] the median time toculture conversion in 79 bedaquiline-treated MDR-TB patients was 83 days; this compares with ourmedian time of 60 days. In the study by DIACON et al. the culture conversion at the end of 24 weeks was79% and at 120 weeks was 58% versus 91.7% at the end of therapy in our study, and the cure rates were58% and 62.4%, respectively [16]. Regarding effectiveness, although difficult to attribute to bedaquiline, wecan report that bedaquiline-containing regimens achieved culture conversion rates >90% at the end oftreatment and treatment success >70% (76.7% when only calculating patients who started their treatmentbefore December 31, 2014), higher than those observed in other MDR-TB cohorts [3, 15, 17].

1.00 Sputum smear conversion

Time days

0.75

0.50

0.25

0

Prop

ortio

n of

pat

ient

s

100 200 400 600 800

Sputum culture conversion

FIGURE 3 Sputum smear and culture conversion rates of 428 culture-confirmed multidrug-resistanttuberculosis patients exposed to bedaquiline-containing regimens.

600

500

400

300

200

100

QTc

F m

s

Baseline Week 4 Week 8 Week 12

FIGURE 2 Median (interquartile range) values of the Friderica-corrected QT interval (QTcF) and its temporaltrends in the cohort in the initial 12 weeks of treatment. Few patients underwent ECG after week 12.

https://doi.org/10.1183/13993003.00387-2017 9

TUBERCULOSIS | S.E. BORISOV ET AL.

PYM et al. [15] conducted a phase 2 trial to assess the safety and efficacy of bedaquiline in 233 patients;culture conversion was seen in 72.2% at 120 weeks, 8.6% of patients discontinued treatment and 6.9% ofpatients died [15].

Regarding safety, in the study by DIACON et al. [16], 13% of patients in the bedaquiline group died (10 outof 79 versus two out of 81 in the placebo group) versus a very similar 33 (13.4%) out of the 247 who hadan evaluable outcome in our cohort.

The most frequent adverse events in our study were nausea (31.5%), otovestibular toxicity (23.3%),peripheral neuropathy (23.3%), vomiting (21.2%), anaemia (20.9%) and arthralgia (20.4%), theirfrequencies being slightly lower than those described during the licensing study by DIACON et al. (41%nausea, 29% vomiting and 37% arthralgia) [17]. Importantly, in the study by DIACON et al. the proportionsof adverse events were similar in the treatment group versus placebo patients, suggesting that they wereprobably due to the background regimen. In this context, other second-line drugs such asfluoroquinolones or clofazimine might contribute to cardiologic or other adverse events [10, 11, 32], andinvite caution and ECG monitoring.

The results of our study demonstrate that overall, bedaquiline-containing regimens achieve a relativelyhigher proportion of treatment success with a relatively lower proportion of adverse events within differentsettings than previously described, although a standardised approach for the notification of the adverseevents was not implemented across the different settings (risk of underreporting).

Of note, culture conversion rates were higher than those reported in cohorts with an analogous degree ofdisease severity, with time to sputum smear and culture conversion identical or earlier to those observedin comparable cohorts; the proportion of treatment success was higher, and the percentage of adverseoutcomes (death or failure) lower than those seen in available study cohorts with matching diseaseseverity; adverse events due to bedaquiline leading to interruption of the drug were relatively uncommon(5.8%) [10, 11].

Our study confirms that bedaquiline-containing regimens are effective, as demonstrated by the fact that asizeable number of patients were treated with salvage regimens due to previous treatment failure,unfavourable resistance profile, toxicity or all three.

The larger group of patients in different nonclinical trial conditions around the globe reinforces previousfindings that bedaquiline is well tolerated and adverse events might be less common than previouslythought.

Enthusiasm over bedaquiline and delamanid has been curtailed following concerns of potentialcardiotoxicity. Both new drugs are associated with QT prolongation, which may potentially lead toarrhythmia and sudden death, a major reason why their combination has not been recommended.Moreover, the new drugs are likely to be associated with a fluoroquinolone and clofazimine, both knownto prolong QT intervals. The specific role of the many drugs with QT-prolonging potential (and theirsummation or synergistic effect) still needs to be fully understood.

QT prolongation occurred in 9.7% of patients; interruption of bedaquiline due to adverse events occurredin 25 (5.8%) patients.

Although information on QT interval was available in 64% of cases, and the timing of their assessmentnot standardised, it seems the majority of cases died for non-heart-related reasons. A more accurateassessment of the adverse events (from data collection to detailed clinical investigation) will allow theevaluation of the relationship between death and QT prolongation. Unfortunately, our data are notstandardised and therefore heterogeneous.

Close monitoring of drug safety should be implemented widely, particularly for rare adverse events. Acomprehensive, population-level pharmacological surveillance in the post-marketing phase might allow abetter assessment of the safety and tolerability profile of bedaquiline, alone or in combination with otherpotentially cardiotoxic anti-TB drugs [45].

The importance of implementing active TB drug-safety monitoring and management when bedaquiline isused programmatically cannot be overemphasised [45].

We underline the importance of using, in future studies, a standardised ECG monitoring protocol allowingto exclude inter- and intra-day variability in QTc measurements. For study purposes, 24-h Holtermonitoring is probably the best strategy to assess the true impact of drugs on QTc. Clearly, this is notfeasible under programmatic conditions.

The strengths of the study are the large cohort, the inclusion of cases from several countries (ranging from7.7% of all patients receiving bedaquiline in the UK to 100% of bedaquiline-treated patients in Argentina,

https://doi.org/10.1183/13993003.00387-2017 10

TUBERCULOSIS | S.E. BORISOV ET AL.

Australia (Victoria State), Greece, Peru, Russian Federation (Arkhangelsk oblast) and Spain) and thedetailed information collected from the participating centres. The large sample size allows the comparisonof treatment outcomes from different settings for the first time.

Some variables, such as the drug-resistance patterns, the number of previous anti-TB treatment cycles andHIV seroprevalence varied between the settings participating in the study.

However, the observational and retrospective design of the study has inbuilt limitations (recent guidelinechanges, different resource settings, different standards of care and dataset differences), so the researchfindings need to be confirmed by larger randomised controlled clinical trials and well-designed prospectiveobservational studies.

Furthermore, several sources of bias may confound the reliability of the findings: missing random selectivityof the patients, heterogeneity between recruited geographic settings, different levels of completeness ofcollected data (especially adverse events and their severity), nonstandardised DST methods for severaldrugs, background regimen variability between countries which differed over time (mainly later generationfluoroquinolones), different TB/HIV co-infection prevalence and availability of antiretroviral drugs.

The effectiveness of bedaquiline might be overestimated, owing to the additive and/or synergistic roleplayed by other effective drugs (e.g. carbapenems, linezolid, etc.) [7, 23–25, 27–31]. New experimentalstudies should be implemented to assess the critical effect of bedaquiline, as well as delamanid, in newcombination therapies.

The different operating procedures adopted in various settings, as well as heterogeneous drug resistancepatterns, may underestimate the real benefits of the bedaquiline-containing regimens.

However, stratified analyses allowed us to better assess the heterogeneity related to time and geographicalfactors. The new information provided by this observational study allows clinicians managingdifficult-to-treat TB cases in programmatic conditions to better understand how to use bedaquiline in casethe minimum number of active drugs necessary to design an effective regimen is lacking [1, 5, 11].

Although new compounds will soon become available to support the move towards TB elimination [46], thisstudy confirms the “core drug” characteristics of bedaquiline in the management of MDR- and XDR-TBcases even in field conditions, and eventually for use in newly designed anti-TB regimens of the future.

While it is crucial to recommend the use of bedaquiline in difficult-to-treat cases, WHO guidelines should befollowed assiduously in order to limit the emergence and spread of bedaquiline-resistant M. tuberculosisstrains.

AcknowledgementsThe authors wish to thank Kerry Uebel (University of the Free State, Bloemfontein, South Africa) for supporting thestudy.

References1 World Health Organization. Global Tuberculosis Report 2016. WHO/HTM/TB/2016.13 Geneva, World Health

Organization, 2016.2 Falzon D, Linh Nhat N, Jaramillo E, et al. The global response to rifampicin-resistant tuberculosis: current

situation and recent trends. Eur Respir J 2016; 48: Suppl. 60, PA1903.3 Migliori GB, Sotgiu G, Gandhi NR, et al. Drug resistance beyond extensively drug-resistant tuberculosis: individual

patient data meta-analysis. Eur Respir J 2013; 42: 169–179.4 Diel R, Vandeputte J, de Vries G, et al. Costs of tuberculosis disease in the European Union: a systematic analysis

and cost calculation. Eur Respir J 2014; 43: 554–565.5 Falzon D, Schünemann HJ, Harausz E, et al. World Health Organization treatment guidelines for drug-resistant

tuberculosis, 2016 update. Eur Respir J 2017; 49: 1602308.6 Shean K, Streicher E, Pieterson E, et al. Drug-associated adverse events and their relationship with outcomes in

patients receiving treatment for extensively drug-resistant tuberculosis in South Africa. PLoS One 2013; 8: e63057.7 Sotgiu G, Pontali E, Migliori GB. Linezolid to treat MDR-/XDR-tuberculosis: available evidence and future

scenarios. Eur Respir J 2015; 45: 25–29.8 Tang S, Yao L, Hao X, et al. Efficacy, safety and tolerability of linezolid for the treatment of XDR-TB: a study in

China. Eur Respir J 2015; 45: 161–170.9 Lee M, Cho SN, Barry CE 3rd, et al. Linezolid for XDR-TB – final study outcomes. N Engl J Med 2015; 373:

290–291.10 Pontali E, Sotgiu G, D’Ambrosio L, et al. Bedaquiline and multidrug-resistant tuberculosis: a systematic and

critical analysis of the evidence. Eur Respir J 2016; 47: 394–402.11 Pontali E, D’Ambrosio L, Centis R, et al. Multidrug-resistance tuberculosis and beyond: an updated analysis of the

current evidence on bedaquiline. Eur Respir J 2017; 49: 1700146.12 Esposito S, D’Ambrosio L, Tadolini M, et al. ERS/WHO Tuberculosis Consilium assistance with extensively

drug-resistant tuberculosis management in a child: case study of compassionate delamanid use. Eur Respir J 2014;44: 811–815.

https://doi.org/10.1183/13993003.00387-2017 11

TUBERCULOSIS | S.E. BORISOV ET AL.

13 Skripconoka V, Danilovits M, Pehme L, et al. Delamanid improves outcomes and reduces mortality inmultidrug-resistant tuberculosis. Eur Respir J 2013; 41: 1393–1400.

14 Gler MT, Skripconoka V, Sanchez-Garavito E, et al. Delamanid for multidrug-resistant pulmonary tuberculosis.N Engl J Med 2012; 366: 2151–2160.

15 Pym AS, Diacon AH, Tang SJ, et al. Bedaquiline in the treatment of multidrug- and extensively drug-resistanttuberculosis. Eur Respir J 2016; 47: 564–574.

16 Diacon AH, Pym A, Grobusch M, et al. The diarylquinoline TMC207 for multidrug-resistant tuberculosis. N EnglJ Med 2009; 360: 2397–2405.

17 Diacon AH, Pym A, Grobusch MP, et al. Multidrug-resistant tuberculosis and culture conversion withbedaquiline. N Engl J Med 2014; 371: 723–732.

18 Guglielmetti L, Le Dû D, Jachym M, et al. Compassionate use of bedaquiline for the treatment ofmultidrug-resistant and extensively drug-resistant tuberculosis: interim analysis of a French cohort. Clin Infect Dis2015; 60: 188–194.

19 Ndjeka N, Conradie F, Schnippel K, et al. Treatment of drug-resistant tuberculosis with bedaquiline in a high HIVprevalence setting: an interim cohort analysis. Int J Tuberc Lung Dis 2015; 19: 979–985.

20 Tiberi S, De Lorenzo S, Centis R, et al. Bedaquiline in MDR/XDR-TB cases: first experience on compassionateuse. Eur Respir J 2014; 43: 289–292.

21 Guglielmetti L, Jaspard M, Le Dû D, et al. Long-term outcome and safety of prolonged bedaquiline treatment formultidrug-resistant tuberculosis. Eur Respir J 2017; 49: 1601799.

22 Tiberi S, D’Ambrosio L, De Lorenzo S, et al. Tuberculosis elimination, patients’ lives and rational use of newdrugs: revisited. Eur Respir J 2016; 47: 664–667.

23 Sotgiu G, Centis R, D’Ambrosio L, et al. Efficacy, safety and tolerability of linezolid containing regimens intreating MDR-TB and XDR-TB: systematic review and meta-analysis. Eur Respir J 2012; 40: 1430–1442.

24 Sotgiu G, Centis R, D’Ambrosio L, et al. Linezolid to treat extensively drug-resistant TB: retrospective data areconfirmed by experimental evidence. Eur Respir J 2013; 42: 288–290.

25 Lee M, Lee J, Carroll MW, et al. Linezolid for treatment of chronic extensively drug-resistant tuberculosis. N EnglJ Med 2012; 367: 1508–1518.

26 Sotgiu G, Centis R, D’Ambrosio L, et al. Low minimal inhibitory concentrations of linezolid againstmultidrug-resistant tuberculosis strains. Eur Respir J 2015; 45: 287–289.

27 Tiberi S, D’Ambrosio L, De Lorenzo S, et al. Ertapenem in the treatment of multidrug-resistant tuberculosis: firstclinical experience. Eur Respir J 2016; 47: 333–336.

28 van Rijn SP, van Altena R, Akkerman OW, et al. Pharmacokinetics of ertapenem in patients withmultidrug-resistant tuberculosis. Eur Respir J 2016; 47: 1229–1234.

29 Tiberi S, Payen MC, Sotgiu G, et al. Effectiveness and safety of meropenem/clavulanate-containing regimens in thetreatment of MDR- and XDR-TB. Eur Respir J 2016; 47: 1235–1243.

30 Tiberi S, Sotgiu G, D’Ambrosio L, et al. Effectiveness and safety of imipenem-clavulanate added to an optimizedbackground regimen (OBR) versus OBR control regimens in the treatment of multidrug-resistant and extensivelydrug-resistant tuberculosis. Clin Infect Dis 2016; 62: 1188–1190.

31 Tiberi S, Sotgiu G, D’Ambrosio L, et al. Comparison of effectiveness and safety of imipenem/clavulanate-versusmeropenem/clavulanate-containing regimens in the treatment of MDR- and XDR-TB. Eur Respir J 2016; 47:1758–1766.

32 Dalcolmo M, Gayoso R, Sotgiu G, et al. Effectiveness and safety of clofazimine within a standard multidrug-resistant tuberculosis regimen in Brazil: first nation-wide report on over 2,500 cases. Eur Respir J 2017; 49:1602445.

33 Tang S, Yao L, Hao X, et al. Clofazimine for the treatment of multidrug-resistant tuberculosis: prospective,multicenter, randomized controlled study in China. Clin Infect Dis 2015; 60: 1361–1367.

34 Van Deun A, Maug AK, Salim MA, et al. Short, highly effective, and inexpensive standardized treatment ofmultidrug-resistant tuberculosis. Am J Respir Crit Care Med 2010; 182: 684–692.

35 Winters N, Butler-Laporte G, Menzies D. Efficacy and safety of World Health Organization group 5 drugs formultidrug-resistant tuberculosis treatment. Eur Respir J 2015; 46: 1461–1470.

36 Alsaad N, Wilffert B, van Altena R, et al. Potential antimicrobial agents for the treatment of multidrug-resistanttuberculosis. Eur Respir J 2014; 43: 884–897.

37 Udwadia ZF, Amale RA, Mullerpattan JB. Initial experience of bedaquiline use in a series of drug-resistanttuberculosis patients from India. Int J Tuberc Lung Dis 2014; 18: 1315–1318.

38 Tadolini M, Lingtsang RD, Tiberi S, et al. First case of extensively drug-resistant tuberculosis treated with bothdelamanid and bedaquiline. Eur Respir J 2016; 48: 935–938.

39 Wallis RS. Cardiac safety of extensively drug-resistant tuberculosis regimens including bedaquiline, delamanid andclofazimine. Eur Respir J 2016; 48: 1526–1527.

40 Tadolini M, Lingtsang RD, Tiberi S, et al. Cardiac safety of extensively drug-resistant tuberculosis regimensincluding bedaquiline, delamanid and clofazimine. Eur Respir J 2016; 48: 1527–1529.

41 Rendon A, Fuentes Z, Torres-Duque CA, et al. Roadmap for tuberculosis elimination in Latin American andCaribbean countries: a strategic alliance. Eur Respir J 2016; 48: 1282–1287.

42 Dalcolmo M, Gayoso R, Sotgiu G, et al. Resistance profile to the drugs composing the ‘shorter’ regimen formultidrug-resistant TB in Brazil, 2000–2015. Eur Respir J 2017; 49: 1602309.

43 Ndjeka N, the Bedaquiline Clinical Access Programme, the Bedaquiline Implementers of the South African NationalTB Programme. Incorporation of Bedaquiline in the South African National TB Programme. 23rd Conference onRetroviruses and Opportunistic Infections (CROI 2016), February 22–25, 2016, Boston. Abstract #754.

44 R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, R Foundation forStatistical Computing, 2006.

45 World Health Organization. Active Tuberculosis Drug-Safety Monitoring and Management (aDSM): Frameworkfor Implementation. WHO/HTM/TB/2015.28. Geneva, World Health Organization, 2015.

46 Lönnroth K, Migliori GB, Abubakar I, et al. Towards tuberculosis elimination: an action framework forlow-incidence countries. Eur Respir J 2015; 45: 928–952.

https://doi.org/10.1183/13993003.00387-2017 12

TUBERCULOSIS | S.E. BORISOV ET AL.